ESTRO 2024 - Abstract Book

S5034

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2024

Frank Brewster 1 , Zoe Middleton 2 , Alan McWilliam 3 , Andrew Brocklehurst 4 , Ganesh Radhakrishna 4 , Robert Chuter 1,3

1 The Christie NHS Foundation Trust, Christie Medical Physics and Engineering, Manchester, United Kingdom. 2 Leeds Teaching Hospitals NHS Trust, Radiotherapy Physics, Leeds, United Kingdom. 3 The University of Manchester, Faculty of Biology, Medicine and Health, Manchester, United Kingdom. 4 The Christie NHS Foundation Trust, Clinical Oncology, Manchester, United Kingdom

Purpose/Objective:

A majority of patients develop hepatocellular carcinoma (HCC) on a background of chronic liver disease, increasing susceptibility to radiation-induced liver disease (RILD). As a result, prescription doses are often limited by liver toxicity, parameterised by liver mean dose (LMD) 1 . Liver function has been shown to be highly heterogeneous due to the irregular anatomical distribution of functioning hepatocytes, especially in HCC patients 2 , but the use of LMD does not take this into account. One possible way to improve treatments for HCC is to take advantage of the heterogeneous liver function by mapping it with functional imaging to reduce the incidence of RILD, avoid reductions in prescription dose or both. This study aimed to demonstrate the feasibility of the use of MR sequences performed without contrast to prioritise the sparing of functional liver. 10 patients were provided for this feasibility study by Perspectum with both T2 and cT1 Hepatica 3 sequences, indicative of fibro-inflammation and hence liver function. Functional images were processed by first downsampling by a factor of 2 in the superior-inferior and left-right planes. They were then segmented using a threshold of 800 ms, below which liver tissue is considered to be healthy. This value was provided by the manufacturer but is validated in studies 4,5 . Finally, the masks were smoothed with a recursive Gaussian filter with a sigma of 4 and thresholded by the median pixel value to provide sparing contours. Plans were then made using the T2 images with population-based density overrides for major tissue groups to allow dose calculation. Plans were made using Monaco (v5.51.11, Elekta AB, Sweden) and the departmental HCC protocol, following NRG-GI003 (NCI-2016-02009), to prescriptions of 30-50 Gy in 5#, limited by liver mean dose. Additionally, OAR tolerances were assessed to 0.5 cm PRVs for the duodenum, small bowel, large bowel, stomach and oesophagus with the PTV cropped back from these PRVs, as per departmental protocol. In addition, sparing plans were made by applying an additional serial planning objective to the functional region only, which was reduced until PTV coverage was compromised. Both plans were reviewed by a clinical oncologist and a second physicist. Dosimetric quantities were extracted for all plans and paired Wilcoxon signed-rank tests were used to test for significant differences between the standard and sparing plans. Material/Methods:

Results:

Clinically acceptable plans for all 10 patients were optimised to meet all mandatory constraints, an example of which is shown in Figure 1. The size of the sparing region varied between patients from 138 to 689 cc, likely due to differing baseline liver function.